[CIG-SHORT] plots of benchmarking results

[Eric Andreas Hetland]

Just one thought, and I may be way off here: but from looking at the
comparisons of the Reverse Slip from the 3 codes to the analytic, it seems
that a large error might be coming from the symmetry BC on the face, and the
reason COMSOL is better is that it does not contain a symmetry BC. Oliver's
notes indicate that he did not use a symmetry BC... It is surprising to be
that the FEs seem to be having trouble with the front symmetry BC on the
reverse slip model, if in fact that is the case here...

- Eric.


On 9/7/06 9:33 PM, "Brad Aagaard" <baagaard at usgs.gov> wrote:

> Hi all:
> 
> I have plots of some of the benchmarking results posted on the
> benchmarking pages. There is a link to the plots page from our main
> benchmark page, 
> http://www.geodynamics.org:8080/cig/workinggroups/eq/workarea/benchmarks/
> 
> I will be making some more plots and finishing up the SCEC poster on
> Fri. I will post more plots as they become available. Also, the ones
> there will probably be replaced with ones with more labels more clearly
> identifying the simulations being compared.
> 
> Conclusions so far...
> 
> * The reverse benchmark results look quite good.
> 
> * The strike-slip benchmark doesn't look as good. It appears that the
> PyLith solution may have an incorrect boundary condition on the bottom
> face of the domain.
> 
> * In both cases, there appears to be a slight difference in the time
> integration between PyLith and GeoFEST. The displacements at t=1 yr have
> greater discrepancies than the displacements at t=0 yr and t=10 yr.
> 
> Brad
> _______________________________________________
> CIG-SHORT mailing list
> CIG-SHORT at geodynamics.org
> http://geodynamics.org/cgi-bin/mailman/listinfo/cig-short
>